A number of different research teams using different approaches have been involved in developing severity indices for use in emergency care evaluation and planning. We believe there is a need to develop a protocol to test the various severity indices in order to: 1) identify the critical strengths and weaknesses of the various approaches to index development, 2) determine the boundaries of use of the various indices, 3) avoid duplication of effort, 4) determine the most productive paths for future research in the severity index field. Because there have been a number of research efforts in severity index development and application and because what is needed is a general protocol applicable to all severity indices, the logical approach to take in developing such a protocol is to involve those who are most knowledgeable in the field--severity index developers and methodologists, experts from the fields of decision analysis and statistics, and users of severity indices from the health care field. The most efficient and economical way to tap these diverse resource people is to bring them together for a working conference. What we are proposing is a conference whose purpose would be to coordinate and integrate the knowledge held by those most experienced in the field and whose end product would be the needed protocol for evaluating severity indices.